> For the complete documentation index, see [llms.txt](https://huataihuang.gitbook.io/cloud-atlas-draft/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://huataihuang.gitbook.io/cloud-atlas-draft/os/linux/storage/filesystem/ext/resize_ext4_root_file_system.md). # EXT4根文件系统resize ## `警告` 本案例能够直接不卸载根文件系统，直接通过fdisk命令修改`/dev/vda3`分区大小，然后`resize2fs`修改EXT4文件系统大小，原因是因为该磁盘恰好只分配了`/dev/sda3`，磁盘后续都是连续空白。 通常情况下，对于EXT4文件系统的底层分区，最好是建立在LVM上，这样才能灵活调整磁盘分区大小。请参考 [扩展LVM](https://github.com/huataihuang/cloud-atlas-draft/tree/6f3204fffc11cf006abd394631e2598d98b415c3/os/linux/storage/storage/device-mapper/lvm/extend_lvm/README.md)。 **如果没有使用LVM，直接通过 fdisk 修改磁盘分区大小，务必要确保分区起始扇区和原先一致，且扩展后的扇区范围没有覆盖其他分区，否则会导致数据丢失。** ``` #fdisk -l /dev/vda Disk /dev/vda: 128.8 GB, 128849018880 bytes, 251658240 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk label type: dos Disk identifier: 0x000cb9dd Device Boot Start End Blocks Id System /dev/vda1 * 2048 2099199 1048576 83 Linux /dev/vda2 2099200 6293503 2097152 82 Linux swap / Solaris /dev/vda3 6293504 125829119 59767808 83 Linux <= 注意：这里vda3没有完全分配完磁盘空间，并且没有后续的vda4，所以才能通过fdisk修改分区大小 ``` ## 操作 * 先删除分许`vda3`然后重新添加`vda3`分区，注意分区起始扇区和原先相同，结束扇区则增大，以便能够使用更多磁盘空间： ``` #fdisk /dev/vda Welcome to fdisk (util-linux 2.23.2). Changes will remain in memory only, until you decide to write them. Be careful before using the write command. Command (m for help): p <= 打印当前磁盘分区 Disk /dev/vda: 128.8 GB, 128849018880 bytes, 251658240 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk label type: dos Disk identifier: 0x000cb9dd Device Boot Start End Blocks Id System /dev/vda1 * 2048 2099199 1048576 83 Linux /dev/vda2 2099200 6293503 2097152 82 Linux swap / Solaris /dev/vda3 6293504 125829119 59767808 83 Linux <= 这个是需要扩展的分区3 Command (m for help): d <= 删除分区 Partition number (1-3, default 3): 3 <= 选择分区3，即删除 Partition 3 is deleted Command (m for help): p <= 再次打印当前磁盘分区 Disk /dev/vda: 128.8 GB, 128849018880 bytes, 251658240 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk label type: dos Disk identifier: 0x000cb9dd Device Boot Start End Blocks Id System /dev/vda1 * 2048 2099199 1048576 83 Linux /dev/vda2 2099200 6293503 2097152 82 Linux swap / Solaris <= 可以看到分区3已经消失 Command (m for help): n <= 添加新分析，也就是再次添加分区3 Partition type: p primary (2 primary, 0 extended, 2 free) e extended Select (default p): p <= 设置新增加的分区是主分区 Partition number (3,4, default 3): 3 <= 分区3 First sector (6293504-251658239, default 6293504): 6293504 <= 关键点：重建的分区3起始扇区必需和原先删除的分区3完全一致 Last sector, +sectors or +size{K,M,G} (6293504-251658239, default 251658239): <= 关键点：重建的分区3的结束扇区值扩大了，完整占据磁盘剩余空间 Using default value 251658239 Partition 3 of type Linux and of size 117 GiB is set Command (m for help): p <= 再次打印当前磁盘分区 Disk /dev/vda: 128.8 GB, 128849018880 bytes, 251658240 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk label type: dos Disk identifier: 0x000cb9dd Device Boot Start End Blocks Id System /dev/vda1 * 2048 2099199 1048576 83 Linux /dev/vda2 2099200 6293503 2097152 82 Linux swap / Solaris /dev/vda3 6293504 251658239 122682368 83 Linux <= 确认重建的分区3正确 Command (m for help): w <= 将分区表信息写回磁盘保存 The partition table has been altered! Calling ioctl() to re-read partition table. WARNING: Re-reading the partition table failed with error 16: Device or resource busy. The kernel still uses the old table. The new table will be used at the next reboot or after you run partprobe(8) or kpartx(8) Syncing disks. ``` * 刷新内核对磁盘分区的识别 注意，由于是修改正挂载的磁盘分区，所以需要通过 `partprobe` 或者 `kpartx` 来通知重新识别。 不过，`partprobe`方式并不能使得内核识别正在使用的根磁盘分区： ``` #partprobe Error: Partition(s) 3 on /dev/vda have been written, but we have been unable to inform the kernel of the change, probably because it/they are in use. As a result, the old partition(s) will remain in use. You should reboot now before making further changes. ``` 检查磁盘分区在内核中信息 ``` cat /proc/partitions | grep vd ``` 显示如下 ``` 253 0 125829120 vda 253 1 1048576 vda1 253 2 2097152 vda2 253 3 59767808 vda3 ``` 可以看到并没有更新成最新的vda3磁盘分区扇区信息。 而`kpartx`是针对Multipath设备，不能使用： ``` #kpartx -v -a /dev/vda device-mapper: reload ioctl on vda1 failed: Invalid argument create/reload failed on vda1 add map vda1 (0:0): 0 2097152 linear /dev/vda 2048 device-mapper: reload ioctl on vda2 failed: Invalid argument create/reload failed on vda2 add map vda2 (0:0): 0 4194304 linear /dev/vda 2099200 device-mapper: reload ioctl on vda3 failed: Invalid argument create/reload failed on vda3 add map vda3 (0:0): 0 245364736 linear /dev/vda 6293504 ``` `partx`设备是针对local设备，注意，不要使用`-v -a`参数，因为不是添加磁盘分区： ``` #partx -v -a /dev/vda partition: none, disk: /dev/vda, lower: 0, upper: 0 /dev/vda: partition table type 'dos' detected partx: /dev/vda: adding partition #1 failed: Device or resource busy partx: /dev/vda: adding partition #2 failed: Device or resource busy partx: /dev/vda: adding partition #3 failed: Device or resource busy partx: /dev/vda: error adding partitions 1-3 ``` 而是应该参考 [Does RHEL 7 support online resize of disk partitions?](https://access.redhat.com/solutions/199573) 使用 `partx` 的 `-u` 参数表示更新： ``` partx -u /dev/vda ``` 此时没有任何输出信息，实际已经更新完成。 再次检查分区大小信息，可以看到已经更新： ``` #cat /proc/partitions | grep vd 253 0 125829120 vda 253 1 1048576 vda1 253 2 2097152 vda2 253 3 122682368 vda3 ``` * `原因解释` * RHEL7内核包含了从[block: add partition resize function to blkpg ioctl](http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c83f6bf98dc1f1a194118b3830706cebbebda8c4)的BLKPG ioctl的修改来支持`BLKPG_RESIZE_PARTITION`操作。 * 当前RHEL7的`util-linux`工具包包含的`partx`和`resizepart`程序是唯一支持`BLKPG_RESIZE_PARTITION`的BLKPG ioctl操作的用户端命令。 * 检查磁盘 ``` tune2fs -l /dev/vda3 ``` * 扩展EXT4文件系统 `resize2fs`命令支持ext2/ext3/ext4文件系统重定义大小。如果文件系统是umount状态，则可以通过`resize2fs`工具扩展或收缩文件系统。如果文件系统是mount状态，则只支持扩展文件系统。注意：要在线扩展文件系统，需要内核和文件系统都支持on-line resize。（现代Linux发行版使用的内核 2.6 可以支持在线resize挂载状态的ext3和ext4；其中，ext3文件系统需要使用`resize_inode`特性） ``` resize2fs [ -fFpPMbs ] [ -d debug-flags ] [ -S RAID-stride ] [ -z undo_file ] device [ size ] ``` 现在我们检查一下当前挂载的`/dev/vda3`磁盘文件系统，挂载为`/`分区，当前大小是`56G`: ``` #df -h Filesystem Size Used Avail Use% Mounted on /dev/vda3 56G 2.4G 51G 5% / ... ``` 执行以下命令扩展文件系统(默认扩展成分区大小，也可以指定文件系统大小) ``` resize2fs /dev/vda3 ``` 显示输出 ``` resize2fs 1.43.5 (04-Aug-2017) Filesystem at /dev/vda3 is mounted on /; on-line resizing required old_desc_blocks = 4, new_desc_blocks = 8 The filesystem on /dev/vda3 is now 30670592 (4k) blocks long. ``` 再次检查挂载的`/`分区，可以看到空间已经扩展到`116G` ``` #df -h Filesystem Size Used Avail Use% Mounted on /dev/vda3 116G 2.4G 108G 3% / ... ``` 此时就可以毫无障碍地使用扩展过的根文件系统。 * 强制系统重启进行fsck RHEL 6等早期使用SysVinit和Debian使用Upstart早期版本，都支持在根分区的文件系统上`/forcefsck`文件来激活强制对根文件系统进行fsck，这是通过`/etc/rc.sysinit`脚本来实现的 ``` touch /forcefsck ``` 这样系统重启会强制进行fack 不过，在systemd系统中，需要通过`systemd-fsck` 服务来设置 [systemd-fsck@.service](https://www.freedesktop.org/software/systemd/man/systemd-fsck@.service.html) 参考 [archliux - fsck](https://wiki.archlinux.org/index.php/fsck) 使用以下命令检查分区设置的fsck检查频率（默认是每30次启动会做一次fsck，不过，当前文件系统设置了 `-1` 强制不检查，或者设置`0`也是不检查） ``` #dumpe2fs -h /dev/vda3 | grep -i "mount count" dumpe2fs 1.43.5 (04-Aug-2017) Mount count: 6 Maximum mount count: -1 ``` 修改检查`/dev/vda3`频率，设置成`1`，则每次重启都会检查： ``` tune2fs -c 1 /dev/vda3 ``` 显示 ``` tune2fs 1.43.5 (04-Aug-2017) Setting maximal mount count to 1 ``` 此时验证文件系统可以看到 ``` #dumpe2fs -h /dev/vda3 | grep -i "mount count" dumpe2fs 1.43.5 (04-Aug-2017) Mount count: 6 Maximum mount count: 1 ``` 现在我们重启操作系统，从VNC终端检查虚拟机可以看到虚拟机启动时进行了文件系统fsck。 既然已经做过fsck了，我们现在恢复原先默认关闭fsck的设置 ``` tune2fs -c -1 /dev/vda3 ``` ## 参考 * [Does RHEL 7 support online resize of disk partitions?](https://access.redhat.com/solutions/199573) - 主要参考 * [Resize a Linux Root Partition Without Rebooting](https://devops.profitbricks.com/tutorials/increase-the-size-of-a-linux-root-partition-without-rebooting/) * [How to resize ext4 root partition live without umount on Linux](https://linuxconfig.org/how-to-resize-ext4-root-partition-live-without-umount) * [6.3. RESIZING AN EXT4 FILE SYSTEM](https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/6/html/storage_administration_guide/ext4grow) - resize2fs命令参考 * [Reload Partition Table Without Reboot In Linux](https://www.teimouri.net/reload-partition-table-without-reboot-linux/#.XEbJLy2B3RY)